rjmccall added a comment. In D76384#1936769 <https://reviews.llvm.org/D76384#1936769>, @mibintc wrote:
> In D76384#1934785 <https://reviews.llvm.org/D76384#1934785>, @rjmccall wrote: > > > Let's make this patch be purely a representation change. I wouldn't expect > > *any* test changes from that; if there are, we should understand why. You > > can then add the stuff about tracking whether there's a pragma in a > > separate patch. > > > OK, some of the tests show ConditionalOperator in the IR matching. Since that > operator is necessarily getting merged into BinaryOperator those tests will > have to change. I'll see about changing the srcpos for the conditional > assignment operator so i can undo the srcpos changes in those tests. I assume you mean CompoundAssignOperator. Sure, if there's some sort of debug output / AST dump in a test that includes expression node names, obviously that's okay to change; and it's possible that the new behavior with CompoundAssignOperator not split out is better in some way. We should just make sure we understand any changes we see. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D76384/new/ https://reviews.llvm.org/D76384 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits